web-dev-qa-db-ja.com

python numpy配列にゼロを埋め込む方法

Numpyバージョン1.5.0でpython 2.6.6を使用して2D numpy配列にゼロを埋め込む方法を知りたい。ごめんなさい!しかし、これらは私の制限です。したがって、np.padは使用できません。たとえば、aをゼロにパディングして、その形状がbに一致するようにします。私がこれをしたい理由は、私ができるようにするためです:

b-a

そのような

>>> a
array([[ 1.,  1.,  1.,  1.,  1.],
       [ 1.,  1.,  1.,  1.,  1.],
       [ 1.,  1.,  1.,  1.,  1.]])
>>> b
array([[ 3.,  3.,  3.,  3.,  3.,  3.],
       [ 3.,  3.,  3.,  3.,  3.,  3.],
       [ 3.,  3.,  3.,  3.,  3.,  3.],
       [ 3.,  3.,  3.,  3.,  3.,  3.]])
>>> c
array([[1, 1, 1, 1, 1, 0],
       [1, 1, 1, 1, 1, 0],
       [1, 1, 1, 1, 1, 0],
       [0, 0, 0, 0, 0, 0]])

私がこれを行うことを考えることができる唯一の方法は、追加ですが、これはかなりいようです。おそらくb.shapeを使用したよりクリーンなソリューションはありますか?

編集、MSeifertsの回答ありがとうございます。私はそれを少しきれいにしなければならなかった、これは私が得たものです:

def pad(array, reference_shape, offsets):
    """
    array: Array to be padded
    reference_shape: Tuple of size of ndarray to create
    offsets: list of offsets (number of elements must be equal to the dimension of the array)
    will throw a ValueError if offsets is too big and the reference_shape cannot handle the offsets
    """

    # Create an array of zeros with the reference shape
    result = np.zeros(reference_shape)
    # Create a list of slices from offset to offset + shape in each dimension
    insertHere = [slice(offsets[dim], offsets[dim] + array.shape[dim]) for dim in range(array.ndim)]
    # Insert the array in the result at the specified offsets
    result[insertHere] = array
    return result
52
user2015487

非常に簡単です。参照形状を使用してゼロを含む配列を作成します。

result = np.zeros(b.shape)
# actually you can also use result = np.zeros_like(b) 
# but that also copies the dtype not only the shape

そして、必要な場所に配列を挿入します。

result[:a.shape[0],:a.shape[1]] = a

そして出来上がり

print(result)
array([[ 1.,  1.,  1.,  1.,  1.,  0.],
       [ 1.,  1.,  1.,  1.,  1.,  0.],
       [ 1.,  1.,  1.,  1.,  1.,  0.],
       [ 0.,  0.,  0.,  0.,  0.,  0.]])

左上の要素を挿入する場所を定義する場合は、少し一般的にすることもできます

result = np.zeros_like(b)
x_offset = 1  # 0 would be what you wanted
y_offset = 1  # 0 in your case
result[x_offset:a.shape[0]+x_offset,y_offset:a.shape[1]+y_offset] = a
result

array([[ 0.,  0.,  0.,  0.,  0.,  0.],
       [ 0.,  1.,  1.,  1.,  1.,  1.],
       [ 0.,  1.,  1.,  1.,  1.,  1.],
       [ 0.,  1.,  1.,  1.,  1.,  1.]])

ただし、許可されているよりも大きなオフセットがないように注意してください。たとえばx_offset = 2の場合、これは失敗します。


任意の数の次元がある場合は、スライスのリストを定義して元の配列を挿入できます。配列と参照が同じ次元数を持ち、オフセットが大きすぎない限り、少し遊んで面白いものを見つけて、任意の形状の配列を(オフセットで)パディングできるパディング関数を作成しました。

def pad(array, reference, offsets):
    """
    array: Array to be padded
    reference: Reference array with the desired shape
    offsets: list of offsets (number of elements must be equal to the dimension of the array)
    """
    # Create an array of zeros with the reference shape
    result = np.zeros(reference.shape)
    # Create a list of slices from offset to offset + shape in each dimension
    insertHere = [slice(offset[dim], offset[dim] + array.shape[dim]) for dim in range(a.ndim)]
    # Insert the array in the result at the specified offsets
    result[insertHere] = a
    return result

そしていくつかのテストケース:

import numpy as np

# 1 Dimension
a = np.ones(2)
b = np.ones(5)
offset = [3]
pad(a, b, offset)

# 3 Dimensions

a = np.ones((3,3,3))
b = np.ones((5,4,3))
offset = [1,0,0]
pad(a, b, offset)
95
MSeifert

NumPy 1.7.0( numpy.pad が追加されたとき)はかなり古い(2013年にリリースされた)ので、質問が方法を求めていたとしてもusing その関数 numpy.pad を使用してどのように達成できるかを知ることが役立つと思いました。

実際には非常に簡単です:

>>> import numpy as np
>>> a = np.array([[ 1.,  1.,  1.,  1.,  1.],
...               [ 1.,  1.,  1.,  1.,  1.],
...               [ 1.,  1.,  1.,  1.,  1.]])
>>> np.pad(a, [(0, 1), (0, 1)], mode='constant')
array([[ 1.,  1.,  1.,  1.,  1.,  0.],
       [ 1.,  1.,  1.,  1.,  1.,  0.],
       [ 1.,  1.,  1.,  1.,  1.,  0.],
       [ 0.,  0.,  0.,  0.,  0.,  0.]])

この場合、0mode='constant'のデフォルト値であることを使用しました。ただし、明示的に渡すことで指定することもできます。

>>> np.pad(a, [(0, 1), (0, 1)], mode='constant', constant_values=0)
array([[ 1.,  1.,  1.,  1.,  1.,  0.],
       [ 1.,  1.,  1.,  1.,  1.,  0.],
       [ 1.,  1.,  1.,  1.,  1.,  0.],
       [ 0.,  0.,  0.,  0.,  0.,  0.]])

2番目の引数([(0, 1), (0, 1)])が紛らわしいように見える場合:各リスト項目(この場合はTuple)はディメンションに対応し、その中の項目はパディングを表しますbefore(最初の要素)および- after(2番目の要素)。そう:

[(0, 1), (0, 1)]
         ^^^^^^------ padding for second dimension
 ^^^^^^-------------- padding for first dimension

  ^------------------ no padding at the beginning of the first axis
     ^--------------- pad with one "value" at the end of the first axis.

この場合、最初の軸と2番目の軸のパディングは同じであるため、2タプルを渡すこともできます。

>>> np.pad(a, (0, 1), mode='constant')
array([[ 1.,  1.,  1.,  1.,  1.,  0.],
       [ 1.,  1.,  1.,  1.,  1.,  0.],
       [ 1.,  1.,  1.,  1.,  1.,  0.],
       [ 0.,  0.,  0.,  0.,  0.,  0.]])

前後のパディングが同じ場合、タプルを省略することもできます(ただしこの場合は適用されません)。

>>> np.pad(a, 1, mode='constant')
array([[ 0.,  0.,  0.,  0.,  0.,  0.,  0.],
       [ 0.,  1.,  1.,  1.,  1.,  1.,  0.],
       [ 0.,  1.,  1.,  1.,  1.,  1.,  0.],
       [ 0.,  1.,  1.,  1.,  1.,  1.,  0.],
       [ 0.,  0.,  0.,  0.,  0.,  0.,  0.]])

または、前後のパディングが同一であるが軸が異なる場合、内側のタプルの2番目の引数を省略することもできます。

>>> np.pad(a, [(1, ), (2, )], mode='constant')
array([[ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.],
       [ 0.,  0.,  1.,  1.,  1.,  1.,  1.,  0.,  0.],
       [ 0.,  0.,  1.,  1.,  1.,  1.,  1.,  0.,  0.],
       [ 0.,  0.,  1.,  1.,  1.,  1.,  1.,  0.,  0.],
       [ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.]])

ただし、間違いを犯しやすい(NumPysの期待があなたの意図と異なる場合)ので、明示的なものを常に使用する傾向があります。

>>> np.pad(a, [1, 2], mode='constant')
array([[ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.],
       [ 0.,  1.,  1.,  1.,  1.,  1.,  0.,  0.],
       [ 0.,  1.,  1.,  1.,  1.,  1.,  0.,  0.],
       [ 0.,  1.,  1.,  1.,  1.,  1.,  0.,  0.],
       [ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.],
       [ 0.,  0.,  0.,  0.,  0.,  0.,  0.,  0.]])

ここで、NumPyは、各軸の前に1つの要素を、すべての軸を2つの要素で埋めたいと考えています!軸1に1つの要素、軸2に2つの要素を埋め込むことを意図していたとしても。

パディングにタプルのリストを使用しましたが、これは単なる「私の慣例」であり、リストのリストやタプルのタプル、さらには配列のタプルも使用できることに注意してください。 NumPyは引数の長さ(または長さがない場合)と各項目の長さ(または長さがある場合)をチェックするだけです!

91
MSeifert

主な問題は、d=b-aを計算する必要があるが、配列のサイズが異なることです。中間のパディングcは不要です

パディングなしでこれを解決できます:

import numpy as np

a = np.array([[ 1.,  1.,  1.,  1.,  1.],
              [ 1.,  1.,  1.,  1.,  1.],
              [ 1.,  1.,  1.,  1.,  1.]])

b = np.array([[ 3.,  3.,  3.,  3.,  3.,  3.],
              [ 3.,  3.,  3.,  3.,  3.,  3.],
              [ 3.,  3.,  3.,  3.,  3.,  3.],
              [ 3.,  3.,  3.,  3.,  3.,  3.]])

d = b.copy()
d[:a.shape[0],:a.shape[1]] -=  a

print d

出力:

[[ 2.  2.  2.  2.  2.  3.]
 [ 2.  2.  2.  2.  2.  3.]
 [ 2.  2.  2.  2.  2.  3.]
 [ 3.  3.  3.  3.  3.  3.]]
6
Juan Leni

配列に1のフェンスを追加する必要がある場合:

>>> mat = np.zeros((4,4), np.int32)
>>> mat
array([[0, 0, 0, 0],
       [0, 0, 0, 0],
       [0, 0, 0, 0],
       [0, 0, 0, 0]])
>>> mat[0,:] = mat[:,0] = mat[:,-1] =  mat[-1,:] = 1
>>> mat
array([[1, 1, 1, 1],
       [1, 0, 0, 1],
       [1, 0, 0, 1],
       [1, 1, 1, 1]])
0
user3409057